From 2b0cf3fed5c0942a31ee871f9a1a7e7c0fc9f783 Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Tue, 7 Oct 2008 18:06:57 -0700
Subject: [PATCH] Fix

---
 src/calibre/ebooks/html.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/src/calibre/ebooks/html.py b/src/calibre/ebooks/html.py
index ea968edd53..f13f5ee2f5 100644
--- a/src/calibre/ebooks/html.py
+++ b/src/calibre/ebooks/html.py
@@ -436,6 +436,7 @@ class Parser(PreProcessor, LoggingInterface):
         ''' Create lxml ElementTree from HTML '''
         self.log_info('\tParsing '+os.sep.join(self.htmlfile.path.split(os.sep)[-3:]))
         src = open(self.htmlfile.path, 'rb').read().decode(self.htmlfile.encoding, 'replace').strip()
+        src = src[src.find('<'):]
         src = self.preprocess(src)
         # lxml chokes on unicode input when it contains encoding declarations
         for pat in ENCODING_PATS: